Solution Brief

The Business Opportunity

¹ Big Data Analytics News:http://bigdataanalyticsnews.com/4-predictions-for-nosql-technologies-in-2016/

To advertise smarter, digital marketers should be able to see how

the data from all devices and platforms (desktop, mobile, tablet,

television, video game console, etc.) is connected to easily target

consumers and communicate with them instantaneously. If a

customer views an ad for a product on her smartphone but

purchases it on her laptop, businesses can now bridge the gap of

conversion the moment it happens.

Smart advertising practices must involve analyzing all data

streams from all devices in order to identify customers’ unique

needs and personalize their buying experiences. Application of

sophisticated machine-learning techniques on graphically stored

data in real time is the right solution for the problem.

Use Case: Cross-Device Marketing

The digital marketing industry is rapidly evolving, both in terms of its

business requirements and the enabling technologies needed to improve

decision-making and gain competitive advantage. While the ad tech

community has effectively utilized NoSQL technology, its concentration

has been on meeting the speed requirements with aggregate-oriented key

value stores and column family databases. To gain and maintain a

competitive advantage in this fast-moving industry, it is crucial for digital

marketing to adopt a massively scalable distributed graph platform that

can handle:

· Graph analytics and real-time relationship discovery

· Integration with the open source stack – Spark, Kafka, HDFS, YARN

· High-speed ingest with parallel querying

· Petabyte-scale to trillions of nodes and edges

An enterprise-grade graph analytics platform can be used to augment

these three broad categories:

· Cross-device marketing

Big data analytics is crucial for evaluating impression opportunities and

brand engagement as the number of internet-connected devices is

exploding. A graph analytics platform lets digital marketers expand

customer reach and more accurately measure advertising effectiveness

by seamlessly connecting the dots between devices to gain a completely

unified view of their target audiences.

· Advanced attribution modeling

With so many ways to reach customers, it can be difficult to attribute

campaign success to the right channel. Statistical averages now fall

short, so using a graph analytics platform to uncover insights deep within

your marketing data from all available data sources is the key to

continually optimizing digital spend and increasing ROI.

· Network intelligence

Cybercrime is a major threat to consumer brands and retailers around the

world, and many security breaches are not identified until the damage

has been done. A graph analytics solution can benefit digital marketers

ThingSpan™:The Distributed Graph Platform for Digital Marketing

by instantly spotlighting any unusual patterns that may be

indicative of bots or malware by correlating data from security and

network solutions with internal and semantic web applications.

Unauthorized activity can then be shut down before financial loss

can occur.

While the value of big data is undeniable, many businesses

struggle with adopting a solution that can help them rapidly

process data to enable immediate decision-making. When they are

not able to realize the value of their data quickly, billions of dollars

may be at stake.

Forrester Research estimates that graph databases will be the fastest-growing area in database management systems, with more

than 25% of enterprises using graph by 2017¹.“

”

Solution Brief

ThingSpan™, Objectivity’s massively scalable graph analytics

platform, is compatible with the distributed open source data

management framework, Hadoop, and utilizes Spark for

high-speed streaming ingest and enriched data processing. It

has the power to transform and analyze real-time advertising

and customer data in context, offering a single logical view

of all data—wherever located—to accelerate parallel

processing. ThingSpan’s subgraph similarity capability helps

find emerging patterns with high degrees of accuracy.

ThingSpan leverages open source tools by supporting the

Hadoop and Spark ecosystem atop a high-performance,

distributed graph platform purpose-built for relationship and

pattern discovery. It runs natively on top of POSIX or HDFS as

a YARN application while using Spark for workflow and data

transformation. It is also designed to support streaming

systems based on Kafka, Flume, and other distributed

messaging tools for streaming data. Integration with other

NoSQL solutions and with Spark via DataFrames allows

ThingSpan to ingest streaming data while maintaining and

persisting relationships as first-class logical models.

This model allows for enriched and transformed data to

simplify the support of complex, multi-dimensional queries

associated with digital marketing applications and analytics.

ThingSpan enables businesses to capture powerful insights

around data relationships to make better-informed decisions

in media buying and campaign management, avoid customer

churn, and discover new ways to reach audiences across all

channels and devices.

The ThingSpan™ Solution

Many organizations rely on big data analytics solutions that involve a data

ingestion layer that processes all data points about customer

data—including ecommerce clickstreams, point-of-sale data, and social

media—before breaking them down into stored memory where they can be

queried. However, while large volumes of data may be processed in this

manner, businesses often rely on micro-batches and may need to wait days

until the relevant data points can be surfaced. In many cases—particularly

when dealing with profit opportunities and customer churn—such delays are

not an option.

In order to generate the high-performance, high-speed processing power and

the sophisticated contextual analysis needed within the digital marketing

industry, an enterprise-class graph analytics solution is essential.

The challenge, however, is finding a distributed graph platform that can

scale to petabytes of data and perform queries in parallel to data ingestion.

While there are many platforms available, few of them offer real-time data

analysis at scale, enabling organizations to capture streaming data and

analyze it in relation to historical and contextual data to gain a 360-degree

view of their customers across a broad array of use cases.

Digital marketers need a highly scalable, real-time graph analytics platform

to analyze massive volumes of advertising and customer data for campaign

effectiveness, business performance, predictive analytics, and other

benchmarks that are high priority to organizations.

The Technical Challenge

Fast Data

Big Data HDFS

Workflow Using DataFrames

Apache Spark

Hadoop

Spark Streaming

Transforms –Filters/Classify/Tag/

Etc..

Transforms –Filters/Classify/Tag/

Etc..

ThingSpanApplications

ThingSpan

ARCHITECTURE DIAGRAM

Objectivity, Inc. delivers massively scalable and highly performant distributed database platforms that are proven to power mission-critical applications for the most demanding and complex datasets in the enterprise. Objectivity helps organizations to rapidly build new Spark Streaming-enabled solutions for finding connections and patterns through graph analytics within petabytes of data, stored in HDFS, to achieve real-time relationship discovery. With a rich history serving Global 1000 customer and partners, Objectivity holds deep domain expertise in fusing vital information from massive data volumes to capture new revenue opportunities, drive competitive advantages and deliver better business value. Objectivity is privately held with headquarters in San Jose, California.Visit http://www.objectivity.com to learn more.

ThingSpan is a massively scalable distributed graph platform designed specifically for the complex issue of extracting actionable insights from Fast Data and Big Data sources to enable real-time relationship discovery. It is architected to integrate with major open source Big Data technologies, such as HDFS and Apache Spark. ThingSpan leverages Objectivity’s core object and graph data-modeling technologies and the company’s rich experience in building fusion solutions.

Main 1 (408) 992-7100 | Fax 1 (408) 992-7171 | Email info@objectivity.comObjectivity, Inc. | 3099 N. First Street, Suite 200 | San Jose, CA 95134